How does the Code Execution skill save tokens?

It processes files locally using Python and only sends metadata summaries back to Claude, rather than passing the entire content of every file into the LLM context.

What is the benefit of using this for 100+ files?

For 100+ files, traditional LLM processing might use 150K+ tokens, whereas this skill completes the same task using roughly 1K tokens, a 99.3% saving.

Is my source code sent to a remote API for processing?

No, the execution happens entirely on your local machine using the provided runtime. Only the resulting metadata or operation summaries are shared with the LLM.

Can I perform Git operations through this skill?

Yes, it includes a dedicated Git API to handle status checks, adding files, and committing changes as part of your automated workflows.

What happens if the execution runtime isn't installed?

You can install it by running the setup.sh script located in the plugin's execution-runtime directory within your local Claude configuration.

Code Execution Engine

Name: Code Execution Engine
Author: mhattingpete

bymhattingpete

•

126

生産性とワークフロー

Executes Python code locally with high-performance API access to achieve up to 99% token savings on bulk repository operations.

概要

The Code Execution skill empowers Claude Code to handle large-scale codebase modifications, audits, and transformations by executing logic locally rather than streaming massive amounts of source code through the LLM. By utilizing specialized APIs for filesystem access, metadata-only code analysis, and batch transformations, it processes hundreds of files in seconds while maintaining a minimal token footprint. This is essential for developers working on large repositories who need to perform complex refactoring, dependency analysis, or git automation without hitting context limits or incurring high API costs.

主な機能

Integrated Git automation for automated staging and commits
Drastic reduction in token consumption (up to 99% for 100+ files)
Batch filesystem operations including search, replace, and bulk moves
Local Python runtime for high-speed file and directory processing
126 GitHub stars
Metadata-only code analysis to minimize context window usage

ユースケース

Performing bulk refactoring or identifier renaming across 50+ files simultaneously
Conducting codebase audits for complexity or dependency patterns across large projects
Automating multi-step file migrations and complex iterative code transformations

概要

主な機能

Integrated Git automation for automated staging and commits
Drastic reduction in token consumption (up to 99% for 100+ files)
Batch filesystem operations including search, replace, and bulk moves
Local Python runtime for high-speed file and directory processing
126 GitHub stars
Metadata-only code analysis to minimize context window usage

ユースケース

Performing bulk refactoring or identifier renaming across 50+ files simultaneously
Conducting codebase audits for complexity or dependency patterns across large projects
Automating multi-step file migrations and complex iterative code transformations